Using AI to transform website data into Notion databases

🌈 Abstract

The article discusses the author's experience with using Notion databases and how they leveraged AI, specifically ChatGPT, to streamline the process of extracting data from a website and transforming it into a Notion database.

🙋 Q&A

[01] Extracting data into a table in ChatGPT

1. What are the three methods the author mentions for extracting data from a website into ChatGPT?

• Copy the web content and paste it directly into ChatGPT
• Save the webpage as an HTML file and import it into ChatGPT
• Print the webpage as a PDF and import it into ChatGPT

2. What challenges did the author face when using the different methods, and how did they address them?

• The author found that the first method (copying and pasting the web content) was the easiest and most straightforward, but it didn't always work as expected, especially with more complex webpages.
• The second method (using HTML) provided more precision, but the author still had to provide specific guidelines to assist ChatGPT when it struggled to extract the information smoothly.
• The debugging process was time-consuming and required the author to speak in a language that computers can understand, as human language alone was sometimes insufficient.

[02] Turning the table into a Notion database

1. How did the author transform the table generated by ChatGPT into a Notion database?

• The author created a new page in Notion, copied the table from ChatGPT, and pasted it into the Notion page.
• The author then clicked on the "..." in the top right corner and selected "Turn into database".
• The author mentioned that there were still some minor edits needed, but the result was good overall.

2. What bonus tip did the author provide for refining the table in Notion? The author mentioned that in Notion, you can drag to fill the same value if you want to add tags to the table.

[03] Takeaways

1. What are the key takeaways the author shared about extracting and organizing data with AI?

• Extracting and organizing data with AI is one of the author's favorite applications, as it saves a lot of time on this tedious task.
• However, the results don't always deliver what the author is hoping for, with missing details or displaying errors.
• Debugging is often necessary, which can be time-consuming and requires some basic knowledge of data structures, an area the author has been gradually learning.